PROJECT SUMMARY Drug resistance and metastasis are both deadly processes in cancer that remain poorly understood. In many instances, resistance or metastasis arise from a small subset of the cells within an individual's tumor that behave differently from the rest. Because these cells are only a small fraction of the cells in a patient's tumor, they cannot be sequenced or profiled using traditional methods. Thus, single-cell analysis provides a window into the variability between cells that underlie these harmful processes; however, methods such as single-cell sequencing are performed after fixing or lysing of the cells, which prevents researchers from being able to perform further downstream analysis such as testing the cells for resistant or invasive phenotypes. These efforts catalogue the molecular variability at the single-cell level, but fail to determine how these variable features relate to the phenotypes present in single cells. My research addresses this hurdle via development of a unbiased, high-throughput sequencing method for identifying variability that is sufficiently ingrained in single cells to generate phenotypes. This approach is based upon Luria and Delbrck's 1943 fluctuation analysis. It combines their clever experimental design with the modern twist of high-throughput sequencing assays. When combined with RNA sequencing, our method (MemorySeq) allows us to quantify gene expression dynamics in order to find single-cell gene expression states that are slowly fluctuating and heritable through multiple cell divisions. We hypothesize that these slowly fluctuating gene expression states allow for significant and ingrained changes in single-cells, which are necessary to generate the detrimental phenotypes of resistance and metastasis in cancer. We aim to use our new MemorySeq method to 1) test the hypothesis that long-lived fluctuations in gene expression underly important phenotypes in cancer, specifically drug resistance and invasion, and to 2) identify transcription factors, kinases, and epigenetic regulator proteins responsible for generating and maintaining these long-lived fluctuations in gene expression. These aims will be accomplished using a highly innovative and complementary approach that combines high-throughput sequencing, CRISPR/Cas9 genetic screening, and single-cell imaging. This line of research will determine the single-cell gene expression signatures of rare resistant and invasive populations in multiple cancer types, and will enumerate the transcription factors, kinases, and epigenetic regulator proteins that govern these expression states. The results of this work will be significant to the cancer research community as they will yield new therapeutic targets to specifically inhibit or destroy these undesirable rare cell populations. Furthermore, this conceptual framework is generalizable and broadly accessible to the scientific research community. In the future, these fluctuation analysis methods can be applied to unravel the contribution of slowly fluctuating gene expression states in other biological processes such as development, wound healing, and cell fate decisions.